what you can do, what GA 3 couldn’t… and vice versa


Google Analytics users must use GA4 by July 1, 2023. Until then, they can continue to work on GA3.

Before the fourth version of Google Analytics becomes mandatory, in 2023, instead of Google Universal, detailed review of the specific features that each version offers.

Metric: Event-based flexible data model VS session-based model

GA4 offers a flexible event-based data model. This type of model is not new. It exists on Mixpanel or Facebook Analytics. With this, GA4 uses tracked events to analyze the user lifecycle.

The engagement rate, which is very important in GA4, uses the indicator of engaged sessions. This includes sessions that lasted at least 10 seconds, recorded at least one conversion event, or counted at least two page or screen views.

Events such as downloading a file or clicking an outbound link are added natively. You can add recommended or custom events. GA4 offers a library of standardized events for this purpose. You can create up to 500 uniquely named events, not including events captured automatically and by advanced metering.

“The benefit of the engagement rate is a more faithful representation of the user’s interactions on the page or mobile application,” explains Maurice Largeron, Google Analytics trainer. “On the other hand, the fact that engagement is based on current factors seems overly flattering to the site owner. It may be useful to set a longer duration, say 30 seconds, to make it more relevant in certain websites. .”

GA3 requires you to configure certain events that are natively honored on GA4, such as: B. File downloads or submitting a form. In fact, GA3 uses a session-based model that aggregates user interactions over a period of time. This model favors bounce rate analysis, which refers to the rate of visitors who have visited a single page of the website. This metric is not only restrictive, but should also be viewed with caution. “It’s a pretty vague criterion, not always fair about it,” said Maurice Largeron. “For example, a user who stays on a page of a website for 10 minutes and exits will be counted with 100% rebound and 0 second visit.”

Reports: VS data segmentation for multiple devices and platforms

GA4 uses multiple identifiers for merging, i.e. merging a data set from different platforms and devices according to identical criteria. This standardizes the measurement across the different sources (web, iOS and mobile). As a result, cross-device and cross-platform insights thrive on GA4. The ability for the user to get an instant answer to questions like: What is the number of users across all platforms? Which platform got the most conversions?…

On GA3, cross-domain tracking must be implemented in code or via Google Tag Manager. You need to add an additional code at the level of each activity you want to track. This makes it difficult to merge website and mobile data on GA3.

Rather than offering reports for multiple devices and platforms, GA3 emphasizes data segmentation, allowing data to be sliced ​​to provide a clearer picture of user actions. Different elements are taken into account: the acquisition source, demographic data, actions on the website… GA3 also allows you to save the created segments in a personal library of available segments.

AI: A tool boosted with AI VS limited automation

GA4 gives pride of place to AI and machine learning. This is an opportunity for instant suggestions and insights. For example, you can get responses to the City dimension with the best performing according to the User dimension. Importantly, machine learning gives you predictions about how likely your users are to opt out or buy.

You can also use the exploration tool to perform cross-platform analysis. Its search engine helps cross variables between website and application. With machine learning capabilities, intelligent alerts also alert you to trend changes.

You can also export your data to BigQuery. With this tool, you can securely store your data in the cloud, combine it with other data, and run queries on any data.

You can also use DebugView to verify that data is displayed correctly in GA4 reports. “DebugView works hand in hand with Google’s tag manager,” analyzes Maurice Largeron. “It allows marketers to correct their markups in real-time in local mode.”

On GA3, you can use a tool like Analytics Intelligence to answer a question like “What products have been purchased more than 100 times in the last month?” If the automation on GA3 is more limited, it is possible to generate reports manually.

Data Processing: Many features with GA4 VS Holes in racket for GA3

Involved in recent cases related to data processing and collection, Google appears to be making the fourth version of its statistics tool its lifeline. Remember that the data collected and processed are considered personal data within the meaning of the GDPR. Your processing is therefore subject to its regulations.

By default, Google Analytics 4 offers automatic IP address anonymization. In order to come closer to the requirements of the GDPR, the user can generate an anonymous user ID and not transmit any personal data to Google Analytics. “We must avoid checking the release of data in the settings,” adds Maurice Largeron. “If this option is enabled, it must be automatically included in the consent level. This must be offered when the user comes to the page.”

GA4 also offers deletion of data at any time and for any reason, deletion of an individual user’s data, or disabling of all or part of data collection. “Google Signal must also not be activated in order to avoid the grouping of user data for the purposes of personalized advertising,” develops Maurice Largeron.

GA3 is less equipped to meet GDPR requirements. To better comply with the GDPR principle of minimization, it may also be useful to review the configuration of your properties and your views. The GDPR provides that “personal data must be adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed”.

In GA3 you can manually request the deletion of data. This function deletes all data for the affected dimension.

Leave a Comment